01 - ggplot2 theme

library(presentresults)
library(ggplot2)

The aim of the themes is to help giving an “IMPACT” touch to the graphs, not to create them. The graph should already be created, i.e. type of plots, and the correct aesthetics, labeling, etc.

The following example aims to plot the percentages of households by category of water sources, by gender of the head of households (This a dummy dataset).

Barplot without formating

data_to_plot <- presentresults::presentresults_MSNA2024_labelled_results_table |>
  dplyr::filter(
    analysis_var == "wash_drinking_water_source_cat",
    group_var == "hoh_gender", 
    group_var_value %in% c("male", "female")
  ) |> 
  dplyr::mutate(label_analysis_var_value = factor(label_analysis_var_value,
                                                  levels = c("Improved",
                                                             "Unimproved",
                                                             "Surface water",
                                                             "Undefined")))

initialplot <- data_to_plot %>%
  ggplot2::ggplot() +
  ggplot2::geom_col(
    ggplot2::aes(
      x = label_analysis_var_value,
      y = stat,
      fill = label_group_var_value
    ),
    position = "dodge"
  ) +
  ggplot2::labs(
    title = stringr::str_wrap(unique(data_to_plot$indicator), 50),
    x = stringr::str_wrap(unique(data_to_plot$label_analysis_var), 50),
    fill = stringr::str_wrap(unique(data_to_plot$label_group_var), 20)
  )

initialplot

Initial plot without theme.

theme_barplot

theme_barplot will give REACH color palette to the bar plot, put the y-axis to 0 to 100%.

initialplot + 
  theme_barplot()

Initial plot without theme_barplot

theme_impact

theme_impact will change the background and color of the title.

initialplot + 
  theme_barplot() +
  theme_impact("reach")

Initial plot without theme_barplot and theme_impact

Other palettes

initialplot + 
  theme_barplot(palette = impact_palettes$impact_palette) +
  theme_impact("impact")

Initial plot without theme_barplot and theme_impact set with IMPACT theme

Some palettes are available in the impact_palettes object.

impact_palettes

$reach_palette
[1] "#58585A" "#EE5859" "#D2CBB8" "#c7c8ca"

$impact_palette
[1] "#000000" "#315975" "#58585A"

$agora_palette
[1] "#581522" "#023C40" "#4F9C35" "#F56741" "#F6ECD0"

$tol_palette
[1] "#322288" "#505050" "#44AA99" "#88CCEE" "#DDCC77" "#EE5859" "#AA4499"
[8] "#721621"

$high_contrast_tol_palette
[1] "#275587" "#CEA936" "#A35464"

$wong_palette
[1] "#EE5859" "#E69F00" "#56B4E9" "#322288" "#F0E442" "#0072B2" "#505050"
[8] "#CC79A7"

$divergent
[1] "#F15B22" "#F58120" "#FBAB35" "#209EA0" "#008083" "#0072B2" "#016060"

$divergent_with_neutral
[1] "#721621" "#D7191C" "#FDAE61" "#FFFFBF" "#97D3C3" "#209EA0" "#016060"

Errors and number of colors

The palette should have enough color to match the scale fill. The impact palette only has 3 colors while the graph needs 4.

data_to_plot <- presentresults::presentresults_MSNA2024_labelled_results_table |>
  dplyr::filter(
    analysis_var == "snfi_fds_cannot_cat",
    group_var == "hoh_gender")

initialplot <- data_to_plot %>%
  ggplot2::ggplot() +
  ggplot2::geom_col(
    ggplot2::aes(
      x = label_analysis_var_value,
      y = stat,
      fill = label_group_var_value
    ),
    position = "dodge"
  ) +
  ggplot2::labs(
    title = stringr::str_wrap(unique(data_to_plot$indicator), 50),
    x = stringr::str_wrap(unique(data_to_plot$label_analysis_var), 50),
    fill = stringr::str_wrap(unique(data_to_plot$label_group_var), 20)
  )
initialplot + 
  theme_barplot(palette = impact_palettes$impact_palette) +
  theme_impact("impact")

Error in `palette()`:
! Insufficient values in manual scale. 4 needed but only 3 provided.

If you need more colors, have a look at the grDevices::colorRampPalette

function_couleur <- grDevices::colorRampPalette(impact_palettes$divergent_with_neutral)
function_couleur(20)

 [1] "#721621" "#91161F" "#B1171D" "#D1181C" "#E1402E" "#ED6F43" "#F99E59"
 [8] "#FDBF74" "#FED892" "#FEF2B0" "#EEF8BF" "#CDEAC0" "#ACDCC2" "#8ACDBF"
[15] "#64BCB4" "#3FABA9" "#1E9A9C" "#148788" "#0A7374" "#016060"

This section comes from this vignette.